Logo Utrecht University



Dr. Hugo Schnack

Dr. Hugo Schnack

Utrecht University


An introduction to machine learning and its applications in language research

Machine learning is a data analysis technique to discover (complex) patterns in datasets related to some property of interest. For example, one may wish to automatically recognize the content of a picture and use that to label or categorize it. These techniques have found their application in many (research) areas, ranging from physics, medicine, social sciences, and, of course, language acquisition. I will give an introduction to what machine learning is, when and how it can (and should) be used, and what it can do – for us. I will take you through some appealing (and appalling!) examples from the literature along with examples from current use and potential applications within language research, including the analysis of vocabulary, speech, ERP, and MRI data to predict an individual’s age or risk of developmental disorders (e.g., dyslexia). It will turn out that, without knowing it, many researchers have already used some kind of machine learning. While machine learning is a powerful tool to extract useful information from data and obtain valuable insights, it should be used with care – like any statistics. Common pitfalls are discussed and I will argue that human intelligence is still necessary to obtain valid and reliable results.

Return to the list of speakers